The Localization And Mapping For Underwater Vehicles Based On Inertial Measurement Units And Sonar

With the development of electronics and computer technology, the technology of underwater vehicles, which is an important tool of underwater exploration, has made rapid progress in recent years. In many unknown high-risk environment, autonomous underwater vehicles (AUV) demonstrate their advantages compared with the manned vehicles. Today, autonomous underwater vehicles have been widely applied in marine survey, oil exploration as well as military field, and showed more and more extensive prospects. In order to achieve the completely autonomous capability, the underwater vehicles should be able to conduct analysis, judgments and decision-making in complex underwater environment so that they can move fast, safely and freely. Among them, the real-time positioning is the key technology to implement the autonomous navigation of underwater vehicles. However, Inertial Navigation System (INS) requires precise initial position, and its positioning error accumulates over time. And we are not able to determine the velocity and pose of the underwater vehicles by Global Positioning System (GPS) because there is no GPS satellite signal underwater. As a result, people began to explore a new autonomous navigation technology, which does not require any prior maps and landmarks. We call it Simultaneous Localization and Mapping (SLAM) technology.The paper first describes the underwater vehicle's main positioning and navigation methods, including Inertial Navigation Systems and GPS/INS integrated navigation and so on, and analyzes their respective strengths and weaknesses. And then the Inertial Navigation System and the SLAM algorithm based on particle filter are introduced detailedly. The PF-SLAM algorithm decompose the full SLAM posterior into vehicle path posterior and landmark posteriors conditioned on the vehicle path estimate. The decomposed posterior can be approximated effectively by particle filter.In the end, the localization and mapping algorithm based on inertial measurement units and sonar for underwater vehicle which is called Inertial-SLAM is introduced in the paper. The Inertial-SLAM algorithm described in this paper is a combination of INS and SLAM. It uses the output data of IMU and the features observed by sonar to estimate the underwater vehicle's position and pose, without using other positioning systems such as GPS. The simulation results of the algorithms and the comparison of EKF-SLAM algorithm and Inertial-SLAM algorithm are given. The feasibility and reliability of Inertial-SLAM algorithm can be proved.